R will force you to understand what you're doing to your data. It is kryptonite to reg monkeys.
Exactly this. OP simian.
In Stata it is embarrisingly simple.
In R, you know, the king of stats, it is a sh*tshow! A literal maze of functions. Eight lines of code for something so simple.
Your fanaticism has convinced me to give R a shot, but I think I'll take a pass. Any stats package that cannot cluster AND have robust std errors is not worth an economist's time.
Keep looking up to the statisticians though, it's a good look.
HAC or Newey West is easy in R. For Cluster, I have no idea
http://lmgtfy.com/?q=R+cluster+robust+standard+errors
First hit, which discusses four options:
http://www.richard-bluhm.com/clustered-ses-in-r-and-stata-2/
I like this approach since it's integrated with making a nice table:
^ Also, note that in R, you can just write a function where you pass
model = y ~ x
data = df
into a function
OLS_robust(model,data)
where you've written out the steps to get the cute regression table inside the function. And then you can stop complaining about, "waaah, so many steps, my hands are sooo tired"
From upthread, the lfe package has solved 99.9% of problems reg monkeys might have. But for whatever reason, old, clunky pre-lfe code comes up first in google searches, so people get stuck thinking this is hard in R.
Its not. Its literally just as easy as Stata. A lot of people want it to be hard because it increases the value of their human capital.
I use R to clean and explore the data. I use STATA to run an print regression output. Say what you will about the flexibility of R, it makes me feel better to know that I'm using a standardized tool to generate my final output (plus as OP mentions, reg options are designed better for economists).
I use R to clean and explore the data. I use STATA to run an print regression output. Say what you will about the flexibility of R, it makes me feel better to know that I'm using a standardized tool to generate my final output (plus as OP mentions, reg options are designed better for economists).
Same.